Science Times: Living Brain to Hand: Bypassing Spinal Cord Quadriplegia Using a Novel Neuroprosthesis

According to a recent report from the Christopher and Dana Reeve Foundation, an estimated 5.6 million people, or about 1.9% of the US population, suffer from paralysis.1 Neuroprosthetic devices offer the possibility of restoring functional movements in quadriplegic and paraplegic patients by supplanting or supplementing the input and output of the nervous system. Brain-computer interfaces, or devices that enable direct communication between the brain and an external computer system, have been a subject of research for decades.2,3Several animal studies paved the path toward human implants, first implemented in the mid-1990s.4-6 First came cochlear implants to restore hearing,7 followed by visual prostheses to improve vision8 and allow visual cursor manipulation.9 Today, the brain-computer interface can be leveraged to create an electronic neural bypass of damaged spinal cord regions and return function to paralyzed limbs.

In a recent study from Ohio State University published in Nature, Bouton et al10 reported exciting results of the first successful real-time reanimation of a quadriplegic patient’s own paralyzed limb with intracortically recorded signals. The patient, a 24-year-old man who suffered a diving accident resulting in a C5/6 level injury, underwent implantation of a high-density microelectrode array overlying the hand area of the left primary motor cortex (Figure, A). He was trained to use his neural bypass system (NBS) 3 times per week for 15 months. His motor cortex activity controlled a neuromuscular electric stimulator (NMES) that directly stimulated his own forearm muscles through a flexible sleeve containing 130 electrodes (Figure, B). A computer monitor was used to portray hand movements for the patient to attempt, and a camera recorded his movements (Figure, C). Chronic recordings from the microelectrode array yielded dozens of single neurons during each session (Figure, D). The neural data were analyzed with the use of wavelet decomposition (Figure, E), an approach that breaks up the data into a series of basis oscillations (wavelets). For example, a musical score can be broken into occurrences and durations of notes (A, A#, B, C, etc). Just as the score can be reconstructed by combining the series of all constituent notes, the neural data can be reconstructed by combining the wavelet series. The authors then looked for relationships between the strength of representation of each wavelet (mean wavelet power) and each of the 6 wrist/hand movements that the patient tried to perform.

Full text access available to all readers.